Accepted Manuscript High-resolution habitat mapping on mud fields: new approach to quantitative mapping of Ocean quahog Artem Isachenko , Yana Gubanova , Alexander Tzetlin , Vadim Mokievsky PII:
S0141-1136(14)00098-1
DOI:
10.1016/j.marenvres.2014.05.005
Reference:
MERE 3894
To appear in:
Marine Environmental Research
Received Date: 4 November 2013 Revised Date:
19 April 2014
Accepted Date: 8 May 2014
Please cite this article as: Isachenko, A., Gubanova, Y., Tzetlin, A., Mokievsky, V., High-resolution habitat mapping on mud fields: new approach to quantitative mapping of Ocean quahog, Marine Environmental Research (2014), doi: 10.1016/j.marenvres.2014.05.005. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT •
We develop method of side-scan sonar data interpretation
•
We create high-resolution map of Arctica islandica settlement by biomass groups
•
We create high-resolution benthic habitat map in homogenous condition
AC C
EP
TE D
M AN U
SC
RI PT
using side-scan sonar
ACCEPTED MANUSCRIPT High-resolution habitat mapping on mud fields: new approach to quantitative mapping of Ocean quahog
Artem Isachenko (1), Yana Gubanova (2), Alexander Tzetlin (1), Vadim Mokievsky (3)
RI PT
(1) MSU, White Sea Biological Station, Russian Federation; (2) MSU, Faculty of Geology, Russian Federation; (3) Shirshov Institute of Oceanology Russian Academy of Sciences, Russian Federation
SC
Corresponding author: Artem Isachenko, [email protected], +79161888396
University, Moscow, Russia
ABSTRACT
M AN U
119234, Leninskiye gory, 1-12, Biological department of Lomonosov Moscow State
During 2009–2012 stocks of the bivalve Arctica islandica (Linnaeus, 1767) (Ocean
TE D
quahog) in Kandalaksha Bay (the White Sea) has been assessed using a side-scan sonar, grab sampling and underwater photo imaging. Structurally uniform localities were highlighted on the basis of side-scan signal. Each type of a signal reflects combination of sediment type, micro topography and structural
EP
characteristics of benthic community. The distribution of A. islandica was the predominant factor in determining community structure. Seabed attributes considered most significant were defined for each type of substrate type. Relations of sonar signal and sediment type
AC C
were used for landscape mapping based on sonar data. Community characteristics at known localities were reliably interpolated to the area of survey using statistical processing of geophysical data.
A method of integrated sonar and sampling data interpretation for high-resolution
mapping of A. islandica by biomass groups, benthic faunal groups and associated habitats was developed. Key words Arctica islandica, habitat mapping, side scan sonar, White sea, macrobenthos, spatial distribution 1. INTRODUCTION
ACCEPTED MANUSCRIPT Arctica islandica (Linnaeus, 1767) – one of the most long living and slow growing species of marine Bivalvia, age of individuals can reach 350 years (Jones, 1983; Schöne et al., 2005). It inhabits subtropical and boreal water depth of 10–150 m (Rowel et al., 1990; Thompson et al., 1980). It is estimated that life span of this species in the White Sea is much lower (Gerasimov, Maksimovich, 1987, 2009). To present time there is evidences
Gorlo and the White Sea Basin (Naumov , 2006).
RI PT
of A. islandica settlements in the Onega, Kandalaskha and Dvina bays, as well as in the
The study area is situated in shallowest inner part of Kandalaksha Bay. This area is one of the rare of the White Sea localities with dense A. islandica settlements, which are
SC
the subject of numerous studies (Brotskaya et al., 1963, Krapivin & Poloskin, 2006, Isachenko et al., 2013).
Maximum depth at the area is 29 m; prevailing depths are 10–14 m. Vysokiy island
M AN U
coastal slope composed mainly with sand or mud and at the depth of 8–10 m fine silts start. All over the study area soft sediments (highly hydrated silts) prevail and hard-bottom areas cover inconsiderable part.
The study area is occupied by A. islandica community. Different polychaetes can be found as subdominant species: Micronephtys minuta, Scoloplos sp., Terebellides stroemi, Alitta virens. A. islandica is a dominant species which define community shape. Therefore
TE D
we assume that distribution of this bivalvian mollusk is a marker of this community type. The main imperative in development of this approach was to create a methodology for mining the reliable information of quantitative settlement characteristics distribution with side-scan sonar data. The main difference in comparison to existing methods of large
EP
scale benthic communities habitat mapping (using side-scan sonar) is it’s adjustment for mesoscale researches.
Currently side-scan sonars have been applied in the research and mapping of
AC C
various biological objects – from distinct populations of individual organisms (mussel and oyster beds, coral reefs , etc.) (Wildish et al., 1998; Smith et al., 1998, 2001; Kvernevik et al., 2002) to the areal mapping of the benthic communities (Freitas et al., 2003). This technique has been used in order to monitor changes in landscapes and biota under the pressure of anthropogenic activity (Kenny & Rees, 1996). Side-scan sonar is a simple and common used method for mapping benthic communities in areas where there are pronounced contrasting features. The mapping of benthic communities using SSS is based on the following principles:
ACCEPTED MANUSCRIPT - The existence of a sufficiently strong correlation between the type of sediment and biota. Major abiotic factors (macro relief, currents, depth) affect sediment distribution at the sea bed as well as biota; - Patches of sediment mosaic and their borders are clearly identified and mapped; - SSS back-scatter characterizes cumulative signal of sediment and macrobenthic
affect bottom micro-relief, altering its acoustic characteristics.
RI PT
fauna, as well as living benthic organisms and biogenic structures which could significantly
The main goal of present study is to investigate the spatial distribution of the dense community in mesoscale (100–1000 m). Necessity of high-resolution mapping of A.
SC
islandica settlement lead to development of novel approach for side-scan sonar data interpretation. 2. METHODS
M AN U
During 2009–2012 side-scan sonar surveys, qualitative sediment sampling and underwater photo imaging were carried out at the test area in Kandalaksha bay, the White Sea.
Within the study area the seafloor is a formation of plains, underwater depressions and banks. The results of SSS survey (2009–2010) provided a detail reconstruction of bottom topography and a spatial distribution of sediment type.
TE D
2.1. Side-scan sonar survey
Side-scan sonar «Hydra» was used to produce bathymetry map of studied polygon and acoustic images of the sea bottom. The side-scan sonar array was permanently attached to the research vessel during the survey.
EP
Initial data processing was performed in real time and included: •recording of sonar echoes;
•recording of navigation data;
AC C
•processing echo-signals;
•visualization of the results after processing. For the survey design 50 meters interval between the sonar profiles was chosen
(Fig. 1). However, due to the complexity of bottom topography it was not possible to maintain this interval throughout the whole survey. In 10% of cases the interval between profiles exceeds 50 m (maximum interval is 70m). As this became apparent during acquisition of the sonar profiles, it was decided to shorten the interval between the profiles to 25-30 m for more dense profiling coverage of the target area. The key factor for the quality of the acquired data was the choice of the profiling direction. Latitudinal location of the profiles was adjusted to the directions of the
ACCEPTED MANUSCRIPT Kandalaksha Bay currents, which facilitated the maintenance of the ships course more accurately. As the result of the survey in 2009–2010 14 sonar profiles were performed (Fig. 1). 2.2. Sampling Sampling grid of survey was designed on the basis of bathymetry data, procured
RI PT
during SSS investigations (2009–2010). During the SSS survey it was investigated, that bottom topography has complex formation, represented by highly patched mosaic of sediment distribution. We assumed, that the spatial distribution of the target species will be related to the heterogeneity of sediment distribution. In this case special spatial design of
SC
sampling was needed. Main aim in distributing sample scheme across the study area was to cover several spatial scales. To achieve this samples were distributed on two transect (SN, WE).
M AN U
Sampling was performed using «Okean» grab (Eleftheriou & McInthyre, 2005) with the sampling area equal to 0.1 m2. A total of 106 samples were collected in 2009–2011 within the study area (Fig. 1). Material was washed on a 0.5 mm mesh size sieve. All macrobenthos species were preserved in 4% formalin and identified to species level where possible (Gaevskaya, 1948; Jirkov, 2001; Naumov, 2006; Starobogatov & Naumov, 1987; Tzetlin et al, 2010). All the species were wet-weighted, A. islandica (as well as other
TE D
Bivalvia) were weighted with shell. 2.3. Underwater imaging
SCUBA based underwater photography was used to observe hard-substrate habitats and small-scale distribution of A. islandica. Images were taken with a Canon 5D Mark 2 in an underwater housing with 2 Inon Z240 strobes. Map of observed sites is provided in Fig.
EP
1. A. islandica population density was assessed with 1 m2 frame temporally placed on bottom. At the same time underwater landscape photographs were made. All the detected
AC C
species were included in habitat description. 2.4. Developing mapping methodology For determining an accurate assessment of A. islandica stocks, a novel method of
SSS data processing was developed. The following biological and geophysical data were taken for interpretation and analysis. 2.4.1. Geophysical characteristics The idea of the proposed interpretation method of SSS data is based on attribute computations within the survey area. Using cluster analysis, a specific group of SSS attributes were assigned to each of three levels of A. islandica density distribution. Each
ACCEPTED MANUSCRIPT level of distribution was provided with geological interpretation in accordance with the data, acquired. Attributes of the sonogram were selected according to following principles: 1. absence of correlation between the sonogram’s attributes (Tab. 1);
sediments. The following attributes were chosen for analysis: 1. depth;
RI PT
2. “a relationship” between sonogram’s attributes and characteristics of the
2. average distribution – characteristics of sediment reflective power: soft sediments
SC
(mud) have low reflectivity, while more solid sediments (sand, gravel) have high reflectance;
3. dispersion – characteristic of sediment homogeneity: high indexes of homogeneity
size fractions;
M AN U
indicates poorly sorted sediments, i.e. containing equal proportions of different particle-
4. distribution asymmetry – index connected with sediment grading level: negative indexes indicates predominance of small particle size fraction, positive indexes indicates predominance of coarse deposit;
5. kurtosis – measure of main sediment fraction predominance;
TE D
6. central frequency – measure of sediments attenuation: low values of index indicates high absorption ability of the sediment; 7. mean distance between peaks of sonogram – complex response of fraction’s spatial distribution;
EP
8. ratio of high and low frequencies pulse – ratio of reflective power of thin and coarse sediments.
For details on how to calculate these attributes please see Gubanova et. al., 2013.
AC C
2.4.2. Biological characteristics
A. islandica biomass distribution were considered as a main biological characteristic. Three levels of A. islandica distribution patterns were identified: a. samples with high biomass of the mollusc (more than 1800 g per m2); b. samples with average biomass (30– 1400 g per m2); samples biomass of A. islandica equal to 0 g per m2. 3.4.3. Algorithm of data interpretation 1. Corresponding biological (sampling) and geophysical data. The GPS positions of sampling stations and sonar profiles were compared for unification of A. islandica distribution and the results of sonar survey. Radius for data unification was chosen equal to 30 m in accordance to accuracy of benthic sampling
ACCEPTED MANUSCRIPT stations positioning. In total, the biomass data from 67 samples were used in analysis, while rest of the samples were excluded from analysis because of far distance from SSS profiles. These samples were used later on to test developed methodology. The number of samples varied inside each level of A. islandica biomass distribution: the level with max.
points, level with no biomass – with 14 points.
RI PT
biomass was presented with 3 sampling points, the level with average biomass - with 50
3. Additional points with A. islandica zero biomass were added to analysis to outline the shallows with no mollusks and with more coarse sediments.
Bottom depressions occupy a significant area at the study site (5–10%), which were
SC
identified from SSS data. These areas at the site are not inhabited with A. islandica; this was confirmed with sampling data and photos taken during the dives. Three points based on the dives (depths of 22-25 m) were added to the total sample database. A summary
M AN U
table with 70 points with known geophysical and biological parameters was formed after these manipulations. A summary table of the rest 3045 points were formed with only data for attributes of the sonar signal (all the tables provided in electronic supplement). 4.Checking the attributes. Selected attributes were tested for correspondence with identified biological clusters, i.e. does sonogram attributes respond to different levels of A. islandica biomass distribution? For such examination pairwise charts of all attributes and
depth factor (Fig. 2).
TE D
biomass were produced. The level with no A. islandica biomass was clearly identified by
The level with a maximum biomass were clearly separated by dispersion and large negative kurtosis values (Fig. 3). The mean values of the variance and kurtosis within the
EP
group with average and maximum values of biomass were different, which should lead to the separation of these groups for further analysis. Привести результаты теста 5. Examine the homogeneity of main parameters (depth, dispersion, kurtosis) of the
AC C
detached groups. Leading parameters of points with high or average biomass were homogenous (Tab. 2), but group with absence of the mollusc clearly separates into two because of depth (Fig. 4). 6. Algorithm of unifying pointed biological and areal spread geophysical data was
based on discriminant analysis. Discriminant functions of all marked biological groups were used for analysis of all points bearing sonar data attributes. Error level was calculated by comparing results of a posteriori and a priori classification. Overall error level for this method was less then 10% (test results see in electronic supplement). SSS mapping provided detailed map of A. islandica settlement biomass distribution as well as outlined contours of homogenous underwater landscape elements (Fig. 5). To
ACCEPTED MANUSCRIPT make complete description of each biotope highlighted underwater imaging during SCUBA diving. 3. RESULTS 3.1. Benthic community The study area is dominated by soft sediments (loosely packed, highly hydrated silt
RI PT
sediments), while hard substrate areas (represented by moraine deposits) have a limited distribution. Bottom current speeds are generally low (0.02 to 0.17 m /sec, the average value – 0.11 m/sec for middle of tidal cycle). Slow current speeds and features of the bottom topography (composition of depressions) form the hydrological conditions in which
organic carbon in surface sediments is 2.75±0.31%.
SC
the off-bottom temperatures are elevated (12-15oC) in the summer months. The content of
An Arctica islandica community generally dominates the study area. 45 invertebrates
M AN U
taxa were observed in the macrozoobenthos: Polychaeta – 27, Crustacea – 8, Bivalvia – 7, Gastropoda, Nemertea, Ascidiacea – 1. The benthos was dominated (both in number and in biomass) by bivalves, particularly A. islandica, and polychaetes. While A. islandica was the dominant species at most stations, other the species with which it occurred changed from station to station but were predominantly polychaetes such as Micronephtys minuta, Alitta virens, Terebellides stroemi and Scoloplos ex gr. acutus.
TE D
The total number of species in samples varied from 5 to 15, average number was 10 species per sample. The number of species and individuals per sample was not correlated with the depth (r= –0.02). Samples with a maximum number of individuals were taken in the central part of the area with varied depths.
EP
Uneven distribution of target species can be explained by the complicity of the underwater topography and high heterogeneity of sediment distribution. The highest A. islandica population density was found on the banks periphery and slopes, such places are the
AC C
most favorable for the settlement not only this, but also other seston feeders as long as it forms good conditions for effective filtration (supply and aeration) due to greater current speed. Large filtration efficiency with increasing flow intensity for various species of bivalves (Walne, 1972).
The average population density is 532 ind./m2, and the average biomass – 419
g/m2. The average biomass at those stations with A. islandica was 495.2 g/m2, and the contribution of A. islandica to the biomass averaged more than 90%. 3.2. Mapping results
ACCEPTED MANUSCRIPT Four types of habitats of underwater landscape were distinguished during SSS interpretation: a. submerged banks; b. mud fields; c. sea-bed depressions (Gubanova, et. al., 2013). These habitats were groundtruthed using photography and grab samples Underwater banks Two types of submerged banks are described within the area. These banks are of
RI PT
the same geological morphology, but at different stages of development.
The first type of bank has ridges which are covered with recent marine sediments (muddy sand, modal fracture = 0.001 mm). Banks depth near ridge is approx. 10 m. Macrobenthos of first type of banks consists of echinoderm Asterias rubens, the gastropod
SC
– Buccinum spp., tunicates and unindentified sponges. These places have maximum biomass of A. islandica. Micronephtys minuta, Terebellides stroemi, Scoloplos acutus, Alitta virens were commonly found as subdominant species.
M AN U
Ridges of the second type submerged banks are more complex by structure on the top and lie at depth less than 6 m. Ridges of these banks are composed by coarse deposit of moraine debris (modal fraction 1.5 cm–1.5 m). These types of banks are very well defined at the SSS images because of high acoustic reflectivity of the coarse material. Large boulders are fouled with sponges (Haliclona aquaeductus as example), kelp (Laminaria saccharina) and other brown sea weeds (i.e. Chorda sp.). The sedentary
TE D
benthos is dominated by the tunicate Styela rustica. Coarse material on the bank ridges is covered with a thin veneer of silt. The large sedentary polychaete Amphitrite figulis which forms small mud lumps are common on the top of the banks. Other macrobenthic species at the bank ridges are the echinoderm Asterias rubens, the hydroid Obelia geniculata, the
EP
gastropod Buccinum spp. and the crustacean Pandalus montagui. A. islandica was not recorded at the tops of this banks. Around the periphery of these banks (5–7 m depth) were covered with aleurite
AC C
(modal fraction 0.01–0.05 mm) and littered with numerous dead shells of A. islandica. Live A. islandica individuals were founded in a very small number (0–1 ind. per m2). Other common macrobenthic species here is the ascidian Styela rustica. Bank slopes (9–12 m depth) are of the same type of the sediment as banks’
periphery. These places have the same maximal density of the A. islandica settlement as first type banks ridges (also the same depth). Mud fields Mud fields (plains) were located in the central and east part of the study area. Mud fields are the most frequently recorded habitat within the study area and are characterized by uniform topography and sediment type (The prevailing sediment type is silt/clay with
ACCEPTED MANUSCRIPT small amounts of sand (
S0141-1136(14)00098-1
DOI:
10.1016/j.marenvres.2014.05.005
Reference:
MERE 3894
To appear in:
Marine Environmental Research
Received Date: 4 November 2013 Revised Date:
19 April 2014
Accepted Date: 8 May 2014
Please cite this article as: Isachenko, A., Gubanova, Y., Tzetlin, A., Mokievsky, V., High-resolution habitat mapping on mud fields: new approach to quantitative mapping of Ocean quahog, Marine Environmental Research (2014), doi: 10.1016/j.marenvres.2014.05.005. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT •
We develop method of side-scan sonar data interpretation
•
We create high-resolution map of Arctica islandica settlement by biomass groups
•
We create high-resolution benthic habitat map in homogenous condition
AC C
EP
TE D
M AN U
SC
RI PT
using side-scan sonar
ACCEPTED MANUSCRIPT High-resolution habitat mapping on mud fields: new approach to quantitative mapping of Ocean quahog
Artem Isachenko (1), Yana Gubanova (2), Alexander Tzetlin (1), Vadim Mokievsky (3)
RI PT
(1) MSU, White Sea Biological Station, Russian Federation; (2) MSU, Faculty of Geology, Russian Federation; (3) Shirshov Institute of Oceanology Russian Academy of Sciences, Russian Federation
SC
Corresponding author: Artem Isachenko, [email protected], +79161888396
University, Moscow, Russia
ABSTRACT
M AN U
119234, Leninskiye gory, 1-12, Biological department of Lomonosov Moscow State
During 2009–2012 stocks of the bivalve Arctica islandica (Linnaeus, 1767) (Ocean
TE D
quahog) in Kandalaksha Bay (the White Sea) has been assessed using a side-scan sonar, grab sampling and underwater photo imaging. Structurally uniform localities were highlighted on the basis of side-scan signal. Each type of a signal reflects combination of sediment type, micro topography and structural
EP
characteristics of benthic community. The distribution of A. islandica was the predominant factor in determining community structure. Seabed attributes considered most significant were defined for each type of substrate type. Relations of sonar signal and sediment type
AC C
were used for landscape mapping based on sonar data. Community characteristics at known localities were reliably interpolated to the area of survey using statistical processing of geophysical data.
A method of integrated sonar and sampling data interpretation for high-resolution
mapping of A. islandica by biomass groups, benthic faunal groups and associated habitats was developed. Key words Arctica islandica, habitat mapping, side scan sonar, White sea, macrobenthos, spatial distribution 1. INTRODUCTION
ACCEPTED MANUSCRIPT Arctica islandica (Linnaeus, 1767) – one of the most long living and slow growing species of marine Bivalvia, age of individuals can reach 350 years (Jones, 1983; Schöne et al., 2005). It inhabits subtropical and boreal water depth of 10–150 m (Rowel et al., 1990; Thompson et al., 1980). It is estimated that life span of this species in the White Sea is much lower (Gerasimov, Maksimovich, 1987, 2009). To present time there is evidences
Gorlo and the White Sea Basin (Naumov , 2006).
RI PT
of A. islandica settlements in the Onega, Kandalaskha and Dvina bays, as well as in the
The study area is situated in shallowest inner part of Kandalaksha Bay. This area is one of the rare of the White Sea localities with dense A. islandica settlements, which are
SC
the subject of numerous studies (Brotskaya et al., 1963, Krapivin & Poloskin, 2006, Isachenko et al., 2013).
Maximum depth at the area is 29 m; prevailing depths are 10–14 m. Vysokiy island
M AN U
coastal slope composed mainly with sand or mud and at the depth of 8–10 m fine silts start. All over the study area soft sediments (highly hydrated silts) prevail and hard-bottom areas cover inconsiderable part.
The study area is occupied by A. islandica community. Different polychaetes can be found as subdominant species: Micronephtys minuta, Scoloplos sp., Terebellides stroemi, Alitta virens. A. islandica is a dominant species which define community shape. Therefore
TE D
we assume that distribution of this bivalvian mollusk is a marker of this community type. The main imperative in development of this approach was to create a methodology for mining the reliable information of quantitative settlement characteristics distribution with side-scan sonar data. The main difference in comparison to existing methods of large
EP
scale benthic communities habitat mapping (using side-scan sonar) is it’s adjustment for mesoscale researches.
Currently side-scan sonars have been applied in the research and mapping of
AC C
various biological objects – from distinct populations of individual organisms (mussel and oyster beds, coral reefs , etc.) (Wildish et al., 1998; Smith et al., 1998, 2001; Kvernevik et al., 2002) to the areal mapping of the benthic communities (Freitas et al., 2003). This technique has been used in order to monitor changes in landscapes and biota under the pressure of anthropogenic activity (Kenny & Rees, 1996). Side-scan sonar is a simple and common used method for mapping benthic communities in areas where there are pronounced contrasting features. The mapping of benthic communities using SSS is based on the following principles:
ACCEPTED MANUSCRIPT - The existence of a sufficiently strong correlation between the type of sediment and biota. Major abiotic factors (macro relief, currents, depth) affect sediment distribution at the sea bed as well as biota; - Patches of sediment mosaic and their borders are clearly identified and mapped; - SSS back-scatter characterizes cumulative signal of sediment and macrobenthic
affect bottom micro-relief, altering its acoustic characteristics.
RI PT
fauna, as well as living benthic organisms and biogenic structures which could significantly
The main goal of present study is to investigate the spatial distribution of the dense community in mesoscale (100–1000 m). Necessity of high-resolution mapping of A.
SC
islandica settlement lead to development of novel approach for side-scan sonar data interpretation. 2. METHODS
M AN U
During 2009–2012 side-scan sonar surveys, qualitative sediment sampling and underwater photo imaging were carried out at the test area in Kandalaksha bay, the White Sea.
Within the study area the seafloor is a formation of plains, underwater depressions and banks. The results of SSS survey (2009–2010) provided a detail reconstruction of bottom topography and a spatial distribution of sediment type.
TE D
2.1. Side-scan sonar survey
Side-scan sonar «Hydra» was used to produce bathymetry map of studied polygon and acoustic images of the sea bottom. The side-scan sonar array was permanently attached to the research vessel during the survey.
EP
Initial data processing was performed in real time and included: •recording of sonar echoes;
•recording of navigation data;
AC C
•processing echo-signals;
•visualization of the results after processing. For the survey design 50 meters interval between the sonar profiles was chosen
(Fig. 1). However, due to the complexity of bottom topography it was not possible to maintain this interval throughout the whole survey. In 10% of cases the interval between profiles exceeds 50 m (maximum interval is 70m). As this became apparent during acquisition of the sonar profiles, it was decided to shorten the interval between the profiles to 25-30 m for more dense profiling coverage of the target area. The key factor for the quality of the acquired data was the choice of the profiling direction. Latitudinal location of the profiles was adjusted to the directions of the
ACCEPTED MANUSCRIPT Kandalaksha Bay currents, which facilitated the maintenance of the ships course more accurately. As the result of the survey in 2009–2010 14 sonar profiles were performed (Fig. 1). 2.2. Sampling Sampling grid of survey was designed on the basis of bathymetry data, procured
RI PT
during SSS investigations (2009–2010). During the SSS survey it was investigated, that bottom topography has complex formation, represented by highly patched mosaic of sediment distribution. We assumed, that the spatial distribution of the target species will be related to the heterogeneity of sediment distribution. In this case special spatial design of
SC
sampling was needed. Main aim in distributing sample scheme across the study area was to cover several spatial scales. To achieve this samples were distributed on two transect (SN, WE).
M AN U
Sampling was performed using «Okean» grab (Eleftheriou & McInthyre, 2005) with the sampling area equal to 0.1 m2. A total of 106 samples were collected in 2009–2011 within the study area (Fig. 1). Material was washed on a 0.5 mm mesh size sieve. All macrobenthos species were preserved in 4% formalin and identified to species level where possible (Gaevskaya, 1948; Jirkov, 2001; Naumov, 2006; Starobogatov & Naumov, 1987; Tzetlin et al, 2010). All the species were wet-weighted, A. islandica (as well as other
TE D
Bivalvia) were weighted with shell. 2.3. Underwater imaging
SCUBA based underwater photography was used to observe hard-substrate habitats and small-scale distribution of A. islandica. Images were taken with a Canon 5D Mark 2 in an underwater housing with 2 Inon Z240 strobes. Map of observed sites is provided in Fig.
EP
1. A. islandica population density was assessed with 1 m2 frame temporally placed on bottom. At the same time underwater landscape photographs were made. All the detected
AC C
species were included in habitat description. 2.4. Developing mapping methodology For determining an accurate assessment of A. islandica stocks, a novel method of
SSS data processing was developed. The following biological and geophysical data were taken for interpretation and analysis. 2.4.1. Geophysical characteristics The idea of the proposed interpretation method of SSS data is based on attribute computations within the survey area. Using cluster analysis, a specific group of SSS attributes were assigned to each of three levels of A. islandica density distribution. Each
ACCEPTED MANUSCRIPT level of distribution was provided with geological interpretation in accordance with the data, acquired. Attributes of the sonogram were selected according to following principles: 1. absence of correlation between the sonogram’s attributes (Tab. 1);
sediments. The following attributes were chosen for analysis: 1. depth;
RI PT
2. “a relationship” between sonogram’s attributes and characteristics of the
2. average distribution – characteristics of sediment reflective power: soft sediments
SC
(mud) have low reflectivity, while more solid sediments (sand, gravel) have high reflectance;
3. dispersion – characteristic of sediment homogeneity: high indexes of homogeneity
size fractions;
M AN U
indicates poorly sorted sediments, i.e. containing equal proportions of different particle-
4. distribution asymmetry – index connected with sediment grading level: negative indexes indicates predominance of small particle size fraction, positive indexes indicates predominance of coarse deposit;
5. kurtosis – measure of main sediment fraction predominance;
TE D
6. central frequency – measure of sediments attenuation: low values of index indicates high absorption ability of the sediment; 7. mean distance between peaks of sonogram – complex response of fraction’s spatial distribution;
EP
8. ratio of high and low frequencies pulse – ratio of reflective power of thin and coarse sediments.
For details on how to calculate these attributes please see Gubanova et. al., 2013.
AC C
2.4.2. Biological characteristics
A. islandica biomass distribution were considered as a main biological characteristic. Three levels of A. islandica distribution patterns were identified: a. samples with high biomass of the mollusc (more than 1800 g per m2); b. samples with average biomass (30– 1400 g per m2); samples biomass of A. islandica equal to 0 g per m2. 3.4.3. Algorithm of data interpretation 1. Corresponding biological (sampling) and geophysical data. The GPS positions of sampling stations and sonar profiles were compared for unification of A. islandica distribution and the results of sonar survey. Radius for data unification was chosen equal to 30 m in accordance to accuracy of benthic sampling
ACCEPTED MANUSCRIPT stations positioning. In total, the biomass data from 67 samples were used in analysis, while rest of the samples were excluded from analysis because of far distance from SSS profiles. These samples were used later on to test developed methodology. The number of samples varied inside each level of A. islandica biomass distribution: the level with max.
points, level with no biomass – with 14 points.
RI PT
biomass was presented with 3 sampling points, the level with average biomass - with 50
3. Additional points with A. islandica zero biomass were added to analysis to outline the shallows with no mollusks and with more coarse sediments.
Bottom depressions occupy a significant area at the study site (5–10%), which were
SC
identified from SSS data. These areas at the site are not inhabited with A. islandica; this was confirmed with sampling data and photos taken during the dives. Three points based on the dives (depths of 22-25 m) were added to the total sample database. A summary
M AN U
table with 70 points with known geophysical and biological parameters was formed after these manipulations. A summary table of the rest 3045 points were formed with only data for attributes of the sonar signal (all the tables provided in electronic supplement). 4.Checking the attributes. Selected attributes were tested for correspondence with identified biological clusters, i.e. does sonogram attributes respond to different levels of A. islandica biomass distribution? For such examination pairwise charts of all attributes and
depth factor (Fig. 2).
TE D
biomass were produced. The level with no A. islandica biomass was clearly identified by
The level with a maximum biomass were clearly separated by dispersion and large negative kurtosis values (Fig. 3). The mean values of the variance and kurtosis within the
EP
group with average and maximum values of biomass were different, which should lead to the separation of these groups for further analysis. Привести результаты теста 5. Examine the homogeneity of main parameters (depth, dispersion, kurtosis) of the
AC C
detached groups. Leading parameters of points with high or average biomass were homogenous (Tab. 2), but group with absence of the mollusc clearly separates into two because of depth (Fig. 4). 6. Algorithm of unifying pointed biological and areal spread geophysical data was
based on discriminant analysis. Discriminant functions of all marked biological groups were used for analysis of all points bearing sonar data attributes. Error level was calculated by comparing results of a posteriori and a priori classification. Overall error level for this method was less then 10% (test results see in electronic supplement). SSS mapping provided detailed map of A. islandica settlement biomass distribution as well as outlined contours of homogenous underwater landscape elements (Fig. 5). To
ACCEPTED MANUSCRIPT make complete description of each biotope highlighted underwater imaging during SCUBA diving. 3. RESULTS 3.1. Benthic community The study area is dominated by soft sediments (loosely packed, highly hydrated silt
RI PT
sediments), while hard substrate areas (represented by moraine deposits) have a limited distribution. Bottom current speeds are generally low (0.02 to 0.17 m /sec, the average value – 0.11 m/sec for middle of tidal cycle). Slow current speeds and features of the bottom topography (composition of depressions) form the hydrological conditions in which
organic carbon in surface sediments is 2.75±0.31%.
SC
the off-bottom temperatures are elevated (12-15oC) in the summer months. The content of
An Arctica islandica community generally dominates the study area. 45 invertebrates
M AN U
taxa were observed in the macrozoobenthos: Polychaeta – 27, Crustacea – 8, Bivalvia – 7, Gastropoda, Nemertea, Ascidiacea – 1. The benthos was dominated (both in number and in biomass) by bivalves, particularly A. islandica, and polychaetes. While A. islandica was the dominant species at most stations, other the species with which it occurred changed from station to station but were predominantly polychaetes such as Micronephtys minuta, Alitta virens, Terebellides stroemi and Scoloplos ex gr. acutus.
TE D
The total number of species in samples varied from 5 to 15, average number was 10 species per sample. The number of species and individuals per sample was not correlated with the depth (r= –0.02). Samples with a maximum number of individuals were taken in the central part of the area with varied depths.
EP
Uneven distribution of target species can be explained by the complicity of the underwater topography and high heterogeneity of sediment distribution. The highest A. islandica population density was found on the banks periphery and slopes, such places are the
AC C
most favorable for the settlement not only this, but also other seston feeders as long as it forms good conditions for effective filtration (supply and aeration) due to greater current speed. Large filtration efficiency with increasing flow intensity for various species of bivalves (Walne, 1972).
The average population density is 532 ind./m2, and the average biomass – 419
g/m2. The average biomass at those stations with A. islandica was 495.2 g/m2, and the contribution of A. islandica to the biomass averaged more than 90%. 3.2. Mapping results
ACCEPTED MANUSCRIPT Four types of habitats of underwater landscape were distinguished during SSS interpretation: a. submerged banks; b. mud fields; c. sea-bed depressions (Gubanova, et. al., 2013). These habitats were groundtruthed using photography and grab samples Underwater banks Two types of submerged banks are described within the area. These banks are of
RI PT
the same geological morphology, but at different stages of development.
The first type of bank has ridges which are covered with recent marine sediments (muddy sand, modal fracture = 0.001 mm). Banks depth near ridge is approx. 10 m. Macrobenthos of first type of banks consists of echinoderm Asterias rubens, the gastropod
SC
– Buccinum spp., tunicates and unindentified sponges. These places have maximum biomass of A. islandica. Micronephtys minuta, Terebellides stroemi, Scoloplos acutus, Alitta virens were commonly found as subdominant species.
M AN U
Ridges of the second type submerged banks are more complex by structure on the top and lie at depth less than 6 m. Ridges of these banks are composed by coarse deposit of moraine debris (modal fraction 1.5 cm–1.5 m). These types of banks are very well defined at the SSS images because of high acoustic reflectivity of the coarse material. Large boulders are fouled with sponges (Haliclona aquaeductus as example), kelp (Laminaria saccharina) and other brown sea weeds (i.e. Chorda sp.). The sedentary
TE D
benthos is dominated by the tunicate Styela rustica. Coarse material on the bank ridges is covered with a thin veneer of silt. The large sedentary polychaete Amphitrite figulis which forms small mud lumps are common on the top of the banks. Other macrobenthic species at the bank ridges are the echinoderm Asterias rubens, the hydroid Obelia geniculata, the
EP
gastropod Buccinum spp. and the crustacean Pandalus montagui. A. islandica was not recorded at the tops of this banks. Around the periphery of these banks (5–7 m depth) were covered with aleurite
AC C
(modal fraction 0.01–0.05 mm) and littered with numerous dead shells of A. islandica. Live A. islandica individuals were founded in a very small number (0–1 ind. per m2). Other common macrobenthic species here is the ascidian Styela rustica. Bank slopes (9–12 m depth) are of the same type of the sediment as banks’
periphery. These places have the same maximal density of the A. islandica settlement as first type banks ridges (also the same depth). Mud fields Mud fields (plains) were located in the central and east part of the study area. Mud fields are the most frequently recorded habitat within the study area and are characterized by uniform topography and sediment type (The prevailing sediment type is silt/clay with
ACCEPTED MANUSCRIPT small amounts of sand (
